Hey guys,
I have finished the MLS, and i feel that there more to learn in the concepts provided, also i feel that by only taking this specialization is not enough for me to have a job in the market.
What are your suggestions ?
Do i have to read books ? Take more courses ? work on projects ?
Also can i start the DLS this way or not ?
Thank you.
@youssef202 having completed MLS I feel you are completely ready to start DLS.
However I can understand your hesitation⌠I would say all of the specializations here are great introductions, but a really through knowledge of the subject does require much more than that, so these are just a start.
Having passed MLS though you are now ready to do more in-depth studies into topics which are of your personal interest. (i.e. am I interested in working with time-series data, with recommender systems, with reinforcement learning, with LLMs, etc ?).
In the end that might not be the final area you get into, but it is best to choose one area that motivates you-- And then from there start working on your own projects.
For a âfinding workâ perspective, I think that is the one thing the specializations here are missing. I did my ML studies elsewhere and at the end, had to complete our own Capstone, which consisted of composing our own individual reports, graphs, code, as well as dealing with unique datasets/problems that had not been broadly covered.
It will be a learning experience for you, but I think having completed such projects is something employers would like to see.
As such I donât think I can recommend one book, without knowing more about your interests.
As to projects, Kaggle.com and the UC Irvine Repository are great places to get ideas/start.
But do take the DLS. I think you will still need to learn more after that, but such is the âstate-of-the-artâ of this subject. However you will learn a whole lot.
Hope that helps.
I hope you really like this because itâs quickly elevated my skills. My method uses ChatGPT4, subscription model, not 4o. I hit the multi-modality headphone icon and begin to speak to it. I ask it to take on the role of a high school professor teaching teenagers about machine learning, computer science, and artificial intelligence in general. I prompt it with something similar to this:
"Ask me 10 questions, one at a time, that all can be answered in an open-ended explaination format. Then, upon me giving the answer, grade it on a percentage basis and suggest in one sentence how I could have given a more complete explaination. "
Guess what!? A couple hours of that and youâll actually have just about mastered a high school level understanding of these concepts. The AI can recognize what would be large chunks of information that might be taught over a year of schooling and chunk it down to major concepts. Then of course, by you needing to verbalize it you learn it better yourself. This was popular in Abraham Lincolnâs time known as âblab schoolsâ where all the kids spoke out the answers to the teacher all at the same time lol.
Okay so, a couple hours in and youâve got more than the basics, youâve grasped a good understanding of it. Guess what? Adjust the prompt to this:
âNow ascend the difficulty and depth of questioning to an undergraduate degree for someone going into computer science. Make your full focus on AI, and ask me 5 questions to ensure I know this level of information in the same format as before.â
Itâll soon be asking you to explain Drop Out, how data augmentation works, why back propogation increases inference speed, what quantizing and vectoring data will do. etc etc.
I stick to around this level, and over weeks, over hundreds of hours, it finally suggested I move to a Masters level of questioning. This has me pretty stumped and requires a lot of explanation and research. I hope my method gives you everything you need!
-Evan Mendenhall
Unfortunately, using a chat tool, you will not know when the information offered is incorrect.
Thank you,
Your answer and opinion is very well understood and beneficial